Custom PMOLED Solutions | Tooling Cost, Shape Options

Tooling for a custom PMOLED is mostly non-recurring engineering, not glass. Semi-custom work that changes only the FPC outline, pin order or firmware usually lands in the low thousands of dollars; a new glass layout with its own photomask set moves into five figures. Shape is the second cost driver, because rectangles cut cheaply and circles do not. Budget the NRE before you negotiate the unit price.

Two kinds of custom

Semi-custom leaves the glass untouched. The factory takes a standard substrate from stock and changes what surrounds it: FPC length and outline, pin assignment, connector style, stiffener and the initialization code inside the driver IC. No photomask is remade, so the schedule is set by FPC fabrication and test-fixture build rather than by the fab.

Full custom starts a new photomask set for the ITO anode, organic layers, cathode and insulator. Cost jumps, and decisions that were fixed in the standard part reopen: driver IC, row and column count, bonding pad layout, encapsulation drawing.

A practical test: if the change can be expressed in Gerber and firmware, treat it as semi-custom. If it needs a mask drawing with an active area, routing area and bonding area, it is full custom.

What you are really paying for

Photomasks dominate. A small monochrome panel typically needs several masks, and the count rises with resolution and with any fine-metal-mask step used to raise pixel density. Aperture tolerance is measured in single-digit micrometres, which is why mask houses charge for precision rather than for area.

Second is FPC tooling: punching and cutting dies, plus a solder-paste stencil if the tail carries decoupling capacitors or a boost circuit. Pitch is the cost lever. A 0.5 mm pitch tail can use conventional dies; 0.3 mm and below usually moves to laser cutting and tighter inspection.

Third is the test fixture. Pogo-pin jigs are built to your pin map, so a 24-pin tail needs 24 probes. Anything non-standard in the driver or in the RAM mapping adds firmware work to rewrite register settings and re-tune gamma.

Fourth is the front end. A cover lens with chamfers and a printed border needs CNC profiling and a screen mesh; adding a projected-capacitive sensor adds an ITO pattern and a tuning pass on the touch controller.

NRE element	Triggered by	Relative weight
Photomask set	New glass layout, resolution or pixel density	Largest
FPC tooling	New outline, pitch below 0.5 mm, SMT on the tail	Moderate
Test jig	New pin count or pad order	Small
Firmware	Non-standard driver, gamma and contrast tuning	Small
Cover lens and touch	CNC profile, printed border, PCAP sensor	Moderate to large

Shape options and glass utilization

Rectangular panels are the cheap case. The layout tiles a mother glass in a regular grid, edge scrap stays narrow, and wheel scribing is fast. Utilization on rectangular monochrome parts is high enough that shape rarely appears as its own line item.

Circular panels are the expensive case. A circle inscribed in a rectangular cell throws away the corners, so utilization falls sharply and the loss is priced into the unit. Cutting also changes: wheels struggle with a continuous arc, so fabs move to UV laser ablation and then polish the edge to remove thermal slag.

Every round display still needs a flat tail for the FPC. That tab is not cosmetic; it carries the fan-out that gathers all pins into the bonding area, and it sets a minimum glass width for the hot-bar or anisotropic-conductive-film bond. Designers who forget the tail discover that the bezel-less circle has a rectangle attached.

Cutouts, dead zones and strength

Holes and irregular outlines add three costs at once: laser time, routing detour and yield. ITO lines that would cross a hole must go around it, which lengthens the trace, raises resistance and can produce a visible brightness gradient across the panel unless the driver compensates.

Encapsulation is the harder constraint. The seal around an arc or a hole must be wider than on a straight edge to preserve the same moisture barrier, and that width becomes a dead zone where no pixel can sit. Interface designers need the CAD drawing before they place text, not after.

Mechanical strength also drops at a re-entrant corner. A common fix is a thin polymer backing film laminated under pressure in a cleanroom, followed by an autoclave cycle to clear bubbles near the cut. Budget the extra lamination step when the outline is complex.

Driving a shaped panel

Passive-matrix OLEDs are addressed row by row, so a pixel is driven well above its average brightness for the duration of its row time and left off for the rest of the frame^[2]. A shaped substrate makes that harder, because traces around the perimeter are longer than traces through the middle.

The result is a resistance difference between centre and edge, which appears as a brightness gradient at high duty cycles. Suppliers manage it by widening traces, splitting the fan-out to two edges, or adding a metal auxiliary line under the ITO. Each option costs mask layers, so ask which one is in the quoted process.

If the dead zone or the pixel count at the corners changes the effective resolution, check any legibility claim under the measurement conditions defined in ISO 9241-307 rather than on a viewbox preview^[1].

Volume, refunds and ownership

Tooling is usually quoted against a minimum order. Semi-custom runs are often a few thousand pieces per batch; full custom is set by how the mother glass is tiled, so the minimum can be higher. Ask for the tiling diagram, not just the MOQ number.

Refund terms vary more than prices. Some suppliers credit the NRE back once cumulative volume passes an agreed threshold; others convert it into a deduction on a later invoice rather than a cash refund. Get the mechanism, the threshold, the time window and the treatment of change orders in writing.

Ownership matters when the shape is part of your industrial design. Paying the NRE in full normally buys the mask and the die, but exclusivity is a separate clause. If you need the tooling destroyed at end of life, say so at contract stage; a certificate and a video of the destruction are cheap to promise and awkward to obtain later.

What to send with the enquiry

Send a drawing, not a description. The package should include the active area and outline with tolerances, the FPC pin map and pitch, the connector type, the driver IC you intend to use, the operating temperature range, the brightness target at a stated duty cycle, and the expected annual volume.

State which parameters are negotiable. Suppliers quote very differently when they know that a 0.2 mm outline change is acceptable but the pin order is fixed by an existing board.

Ask for a sample schedule with the mask and FPC steps separated. That breakdown shows whether the lead time is real engineering or queue time, and it gives you a checkpoint to manage.

Keeping a second source open

A custom shape tends to become single-sourced, because the mask and the die belong to one factory. That is acceptable if you plan for it: keep the FPC on a standard pitch, document the initialization table, and avoid a driver that only one supplier can tune.

Standard PMOLED display modules and graphic OLED modules can carry a first product generation while a custom glass is in tooling. That staging keeps the schedule honest and gives the mask design a chance to settle.

If the shape is a circle, compare the tooling quote against a standard round screen before assuming a custom substrate is required.

Custom PMOLED tooling is a short list of understood items: masks, FPC dies, a test jig, firmware and the front end. Decide whether you are changing glass or only the tail, then price the shape honestly. The cheapest quote is rarely the one with the shortest path to a second source.

References

1. ISO 9241-307:2008 — Ergonomics of human-system interaction — Part 307: Analysis and compliance test methods for electronic visual displays. International Organization for Standardization. https://www.iso.org/standard/40102.html
2. Hilsum, C. (2010). Flat-panel electronic displays: a triumph of physics, chemistry and engineering. Philosophical Transactions of the Royal Society A, 368(1914), 1027–1082. https://doi.org/10.1098/rsta.2009.0247